A quadratic measure of dependence
Statistics Theory
2007-06-13 v1 Statistics Theory
Abstract
Asymptotic properties of a dimension-robust dependence measure are investigated. It is related to those used in independence tests, but is derivable, thus suitable for independent component analysis. An adjustable kernel allows to accelerate the convergence of the estimator without affecting the bias.
Cite
@article{arxiv.math/0609259,
title = {A quadratic measure of dependence},
author = {Sophie Achard},
journal= {arXiv preprint arXiv:math/0609259},
year = {2007}
}
Comments
18 pages, 3 figures, uses natbib.sty